Repeatability
High
The transformation rules are consistent: convert units to metric, convert prices to USD, and map vendor codes to internal SKUs. This same pipeline can be reused every time a new vendor price list arrives.
Ambiguity Tolerance
High
Success criteria are concrete and verifiable: all units in metric, all prices in USD, and a complete vendor_code-to-SKU mapping table. An agent can validate its own output against these rules programmatically.
Data & Tool Availability
Medium
The agent needs the PDF, the existing spreadsheet, and access to a live currency exchange API — all obtainable. PDF table extraction can be lossy depending on formatting, which may require a fallback parsing step.
Error Cost
Medium
Incorrect unit conversions or wrong currency rates would corrupt pricing data, which could affect purchasing decisions or customer-facing prices. However, the output is a spreadsheet that can be reviewed before import, making errors reversible if caught.
Human Judgment Required
Low
The SKU mapping is the only step requiring judgment — matching vendor codes to internal SKUs may involve ambiguous or partial matches. A human should review low-confidence matches, but the bulk of the work is deterministic.